From mboxrd@z Thu Jan 1 00:00:00 1970 Received: from mail.bootlin.com ([62.4.15.54]) by bombadil.infradead.org with esmtp (Exim 4.90_1 #2 (Red Hat Linux)) id 1fXiKw-0007vT-0I for linux-mtd@lists.infradead.org; Tue, 26 Jun 2018 07:24:27 +0000 Date: Tue, 26 Jun 2018 09:24:13 +0200 From: Boris Brezillon To: Miquel Raynal Cc: Martin Kaiser , David Woodhouse , Sascha Hauer , Fabio Estevam , linux-mtd@lists.infradead.org, linux-kernel@vger.kernel.org Subject: Re: [PATCH] mtd: rawnand: mxc: remove __init qualifier from mxcnd_probe_dt Message-ID: <20180626092413.703249c8@bbrezillon> In-Reply-To: <20180626091005.08d56b2a@xps13> References: <1529679025-5735-1-git-send-email-martin@kaiser.cx> <20180626091005.08d56b2a@xps13> MIME-Version: 1.0 Content-Type: text/plain; charset=UTF-8 Content-Transfer-Encoding: quoted-printable List-Id: Linux MTD discussion mailing list List-Unsubscribe: , List-Archive: List-Post: List-Help: List-Subscribe: , On Tue, 26 Jun 2018 09:10:05 +0200 Miquel Raynal wrote: > Hi Martin, Boris, >=20 > On Fri, 22 Jun 2018 16:50:25 +0200, Martin Kaiser > wrote: >=20 > > mxcnd_probe_dt is called from mxcnd_probe, which is not marked as __ini= t. > > =20 >=20 > I think this line is just a side note and should be at the end of the > commit log. >=20 > > Using the sysfs unbind, bind nodes, mxcnd_probe and mxcnd_probe_dt can > > potentially be called at any time. After the __init functions are clean= ed, > > mxcnd_probe_dt is no longer available. Calling it anyway causes a crash= . =20 >=20 > The above paragraph explains the problem and the solution, you can just > add something like "mcvnd_probe is untouched because it is already not > marked as __init" instead of the first line. >=20 > >=20 > > Signed-off-by: Martin Kaiser =20 >=20 > Boris, do you think this patch is a good candidate for stable? >=20 > If yes, Martin, could you please add a couple of stable/fixes tags above > your SoB? We should at least have a Fixes tag. For the stable one, I'm not so sure because the bug does not exist in practice (the compiler always inline mxcnd_probe_dt()). >=20 > Thanks, > Miqu=C3=A8l >=20 > > --- > > I hope that my understanding is correct. > >=20 > > In practice, I could not get the kernel to crash, my compiler would > > inline mxcnd_probe_dt into mxcnd_probe and ignore the __init. > >=20 > > If I marked mxcnd_probe_dt as noinline, I could trigger a crash as > > follows > >=20 > > [root@host ]# echo bb000000.nand > /sys/bus/platform/drivers/mxc_nand/u= nbind > > [root@host ]# echo bb000000.nand > /sys/bus/platform/drivers/mxc_nand/b= ind > > Internal error: Oops - undefined instruction: 0 [#1] ARM > > ... > >=20 > > drivers/mtd/nand/raw/mxc_nand.c | 4 ++-- > > 1 file changed, 2 insertions(+), 2 deletions(-) > >=20 > > diff --git a/drivers/mtd/nand/raw/mxc_nand.c b/drivers/mtd/nand/raw/mxc= _nand.c > > index 26cef21..90cfb5e 100644 > > --- a/drivers/mtd/nand/raw/mxc_nand.c > > +++ b/drivers/mtd/nand/raw/mxc_nand.c > > @@ -1686,7 +1686,7 @@ static const struct of_device_id mxcnd_dt_ids[] = =3D { > > }; > > MODULE_DEVICE_TABLE(of, mxcnd_dt_ids); > > =20 > > -static int __init mxcnd_probe_dt(struct mxc_nand_host *host) > > +static int mxcnd_probe_dt(struct mxc_nand_host *host) > > { > > struct device_node *np =3D host->dev->of_node; > > const struct of_device_id *of_id =3D > > @@ -1700,7 +1700,7 @@ static int __init mxcnd_probe_dt(struct mxc_nand_= host *host) > > return 0; > > } > > #else > > -static int __init mxcnd_probe_dt(struct mxc_nand_host *host) > > +static int mxcnd_probe_dt(struct mxc_nand_host *host) > > { > > return 1; > > } =20